Training Method for Artificial Intelligence Communication Tool and Electronic Device

This application claims the priority benefit of Taiwan application serial no. 113142878, filed on November 8, 2024. The entirety of the above-mentioned patent application is hereby incorporated by reference herein and made a part of this specification.

The present disclosure relates to a training method and an electronic device, and particularly relates to a training method for an artificial intelligence (AI) communication tool and an electronic device.

Generally speaking, for responses automatically generated by an artificial intelligence (AI) feedback system, users may select one or more dissatisfaction options with the responses, or provide additional feedback corresponding to the responses. However, the feedback formats provided by AI feedback systems are typically overly simple, and the formats of additional feedback that users may provide are excessively limited.

In order for an AI feedback system to generate ideal responses, a large quantity of response samples is typically required. Since deliberate response samples require substantial time to produce, most users lack the willingness to provide response samples. Generally speaking, users are more inclined to provide suggestions corresponding to responses through the selection of checkboxes. However, the checkbox method fails to fully express user’s feedback, and development teams find it difficult to construct complete response samples based solely on the checkboxes selected by users.

The present disclosure provides an electronic device and a training method for an artificial intelligence (AI) communication tool, which may more quickly provide users with ideal communication results.

In an embodiment of the present disclosure, a training method is provided for AI communication tools. The AI communication tool is installed in a computer system, and the computer system includes a processor. The processor is configured to execute the training method. The training method includes: by the AI communication tool, providing a communication result and receiving a correction prompt corresponding to the communication result; by a prompt analysis model, generating at least one answer based on the correction prompt and generating at least one similarity between the correction prompt and the at least one answer based on the at least one answer; and by the prompt analysis model, selecting an optimized answer from the at least one answer based on the at least one similarity, correcting the communication result based on the optimized answer to generate a corrected communication result, and providing the corrected communication result to the AI communication tool.

In an embodiment of the present disclosure, the electronic device includes an operation interface and a processor. The processor is coupled to the operation interface. The processor includes the AI communication tool and the prompt analysis model. The processor is operated based on the above training method.

Based on the above, the prompt analysis model automatically generates at least one answer according to the correction prompt. The at least one answer involves different response samples used to train the AI communication tool. In this way, users do not need to spend so much time formulating multiple response samples. The training method may obtain a large amount of response samples. In addition, the prompt analysis model selects an optimized answer from the at least one answer and corrects the communication result according to the optimized answer to generate a corrected communication result. The training method may also generate a corrected communication result that is the closest to the correction prompt.

1 FIG. 1 FIG. 100 110 120 120 121 122 121 122 121 122 Please refer to,illustrates a schematic view of an electronic device according to an embodiment of the present disclosure. In this embodiment, the electronic deviceincludes an operation interfaceand a processor. The processoris configured to execute an artificial intelligence (AI) communication tooland a prompt analysis model. The AI communication tooland the prompt analysis modelmay be implemented through one of computing circuits or any combination thereof, and this disclosure does not limit the implementation methods of the AI communication tooland the prompt analysis model.

121 121 121 When in use, the AI communication toolmay output a corresponding communication result CS according to data and questions provided by the user. For example, the "question" is, for example, a question regarding data content or a question regarding definitions of technical terms, etc. The AI communication tooltakes into consideration the data and the content of the question, and outputs the communication result CS that is logical and exclude erroneous information as much as possible. For example, when the user's question is about the definition of a technical term (for example, what is a green partner), since the question does not directly provide relevant information of the technical term, the communication result CS output by the AI communication toolmay be different from the output expected by the user, or the explanation of the technical term includes erroneous information. When the user is not satisfied with the communication result CS, the user may provide a correction prompt CP corresponding to the communication result CS.

122 1 1 1 122 1 122 1 1 122 2 2 In this embodiment, the prompt analysis modelreceives the correction prompt CP provided by the user and generates answers ANS~ANSn according to the correction prompt CP, and generates similarities SML~SMLn between the correction prompt CP and the answers ANS~ANSn. The prompt analysis modelselects an optimized answer ANS according to the similarities SML~SMLn. For example, the prompt analysis modelcompares the correction prompt CP with the answer ANSto generate the similarity SML. The prompt analysis modelcompares the correction prompt CP with the answer ANSto generate the similarity SML, and so on.

1 122 1 122 1 1 Taking the answer ANSas an example, the prompt analysis modelretrieves at least one first message from the answer ANSand retrieves at least one second message from the correction prompt CP. The prompt analysis modelgenerates the similarity SMLcorresponding to the answer ANSaccording to at least one of a text similarity, a factual relationship, and a contradictory relationship between the at least one first message and the at least one second message.

122 1 1 122 122 121 In this embodiment, the prompt analysis modelselects the optimized answer ANS from the answers ANS~ANSn according to the extent of the similarities SML~SMLn. In addition, the prompt analysis modelcorrects the communication result CS according to the optimized answer ANS to generate a corrected communication result CS'. Furthermore, the prompt analysis modelprovides the corrected communication result CS' to the AI communication tool.

121 130 1 110 110 1 130 In this embodiment, the AI communication toolmay submit the communication result CS and the corrected communication result CS' to a databaseas training data for subsequent training. In addition, the user may edit the corrected communication result CS' to generate and submit the edited corrected communication result CS'' to the operation interface. The operation interfacemay submit at least one of the communication result CS and the edited corrected communication result CS'' to the databaseas training data for subsequent training.

120 In this embodiment, the processoris, for example, a Central Processing Unit (CPU), or other programmable general-purpose or special-purpose microprocessor, a Digital Signal Processor (DSP), a programmable controller, an Application Specific Integrated Circuit (ASIC), a Programmable Logic Device (PLD) or other similar device or combinations of these devices.

121 122 In this embodiment, the AI communication tooland the prompt analysis modelare deep learning models such as a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN), respectively, but the present disclosure is not limited thereto.

100 In this embodiment, the electronic devicemay be a smartphone, a tablet computer, a laptop computer, a personal computer or a portable electronic device, but the present disclosure is not limited thereto.

1 FIG. 2 FIG. 2 FIG. 100 100 100 110 130 110 121 110 110 Please refer toand,illustrates a flowchart of a training method according to an embodiment of the present disclosure. In this embodiment, a training method Smay be applicable to the electronic device. The training method Sincludes steps S~S. In step S, the AI communication toolprovides the communication result CS corresponding to a question inputted by a user. In step S, when the user is not satisfied with the communication result CS, the user may input the correction prompt CP corresponding to the communication result CS through the operation interface. The correction prompt CP is, for example, to correct specific content in the communication result CS, or to correct the definition of a technical term in the communication result CS.

120 122 1 1 1 1 1 122 1 1 1 In step S, the prompt analysis modelgenerates the answers ANS~ANSn according to the correction prompt CP and generates the similarities SML~SMLn between the correction prompt CP and the answers ANS~ANSn. The answers ANS~ANSn are, for example, generated by a Large Language Model (LLM), and the answers ANS~ANSn may include information that does not match actual conditions or content that does not conform to common sense. The prompt analysis modelmay compare the content of the correction prompt CP with the answers ANS~ANSn, thereby acquiring the extent of similarity between the answers ANS~ANSn and the correction prompt CP to generate the corresponding similarities SML~SMLn.

130 122 1 1 In step S, the prompt analysis modelselects the optimized answer ANS from the answers ANS~ANSn according to the similarities SML~SMLn and corrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS'.

1 1 1 1 1 1 1 1 122 1 1 122 122 In this embodiment, the similarities SML~SMLn are, for example, represented by numerical values. The similarities SML~SMLn represent the extent of semantic proximity between the answers ANS~ANSn and the correction prompt CP as well as the extent of content similarity between the answers ANS~ANSn and the established facts. The extent of semantic proximity is, for example, a semantic distance between different texts that is measured based on cosine similarity. For example, different words with a small semantic gap have a close semantic distance. Different words with a large semantic gap have a farther semantic distance. For example, the smaller the difference between the answer ANSand the correction prompt CP is, the higher the numerical value of the similarity SMLis. The larger the difference between the answer ANSand the correction prompt CP is, the lower the numerical value of the similarity SMLis. Therefore, the prompt analysis modelselects the optimized answer ANS from the answers ANS~ANSn according to the extent of the similarities SML~SMLn. Furthermore, the prompt analysis modelcorrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS'. In this embodiment, the optimized answer ANS is the answer corresponding to the highest similarity (i.e., the highest score). Therefore, the corrected communication result CS' is the result closest to the correction prompt CP generated by the prompt analysis model.

122 1 1 121 100 122 1 It is worth mentioning here that the prompt analysis modelautomatically generates the answers ANS~ANSn according to the correction prompt CP. The answers ANS~ANSn may be utilized as response samples for training the AI communication tool. In this way, the user does not need to spend so much time to submit multiple response samples. The training method Smay obtain a large amount of response samples. Furthermore, the prompt analysis modelselects the optimized answer ANS from the answers ANS~ANSn and corrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS'. The training method may also generate the corrected communication result CS' that is the closest to the correction prompt CP.

1 FIG. 3 FIG.A 3 FIG.B 3 FIG.A 3 FIG.B 200 100 200 201 213 201 110 121 Please refer to,and.andillustrate flowcharts of a training method according to an embodiment of the present disclosure. In this embodiment, a training method Sis applicable to the electronic device. The training method Sincludes steps S~S. In step S, the user inputs a message on the operation interface, and the AI communication toolgenerates the communication result CS according to the message.

202 121 121 121 121 130 203 In step S, the AI communication toolreceives feedback from the user. When the feedback is positive, it means that the user is satisfied with the communication result CS generated by the AI communication tool. That is to say, the communication result CS generated by the AI communication toolis accurate. Therefore, the AI communication toolsubmits the message and the communication result CS that satisfies the user to the databasein step S.

110 121 204 205 122 1 1 1 206 122 1 1 122 206 When the feedback is negative, it means that the user is not satisfied with the communication result CS, and the user may input the correction prompt CP on the operation interface. The correction prompt CP is, for example, a suggestion for how the communication result CS may be improved. Therefore, the AI communication toolreceives the correction prompt CP corresponding to the communication result CS in step S. In step S, the prompt analysis modelgenerates the answers ANS~ANSn according to the correction prompt CP and generates the similarities SML~SMLn between the correction prompt CP and the answers ANS~ANSn. In step S, the prompt analysis modelselects the optimized answer ANS from the answers ANS~ANSn according to the extent of the similarities SML~SMLn. Furthermore, the prompt analysis modelcorrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS' in step S.

207 121 In step S, the AI communication toolreceives feedback from the user and classifies the corrected communication result CS' according to the user's feedback.

121 121 208 121 121 130 212 207 When the feedback is positive, it means that the user is satisfied with the corrected communication result CS' generated by the AI communication tool. That is to say, the corrected communication result CS' generated by the AI communication toolis accurate. Therefore, in step S, the AI communication toolclassifies the corrected communication result CS' as a first classification according to the user's positive feedback. Next, the AI communication toolsubmits the corrected communication result CS' classified as the first classification to the databasein step S. Then, the AI communication tool may return to step Sto continue receiving other feedback.

209 121 On the other hand, in step S, when the feedback is negative, it means that the user is not satisfied with the corrected communication result CS', and the AI communication tooldetermines whether the user chooses to edit the corrected communication result CS' according to the user's negative feedback.

110 121 210 121 130 212 When the user is not satisfied with the content of the corrected communication result CS' and does not want to edit the corrected communication result CS' through the operation interface, the AI communication toolclassifies the corrected communication result CS' as a second classification according to the user's negative feedback in step S. Next, the AI communication toolsubmits the corrected communication result CS' to the databasein step S.

110 110 1 211 121 1 213 121 1 130 110 1 130 207 When the user is not satisfied with the content of the corrected communication result CS' and wants to edit the corrected communication result CS' through the operation interface, the user may edit the corrected communication result CS' on the operation interfaceto generate the edited corrected communication result CS''. Therefore, in step S, the AI communication toolreceives the edited corrected communication result CS''. In step S, the AI communication toolsubmits the edited corrected communication result CS'' to the database. For example, the user may edit all the content of the corrected communication result CS', or only correct part of the content of the corrected communication result CS'. The user clicks the submit button on the operation interface, thereby submitting the edited corrected communication result CS'' to the database. Then, the AI communication tool may return to step Sto continue receiving other feedback.

1 122 213 In some embodiments, the edited corrected communication result CS'' may also be submitted to the prompt analysis modelin step S.

1 FIG. 3 FIG.A 3 FIG.B 4 FIG. 4 FIG. 300 300 310 360 310 130 130 Please refer to,,, and.illustrates a flowchart of a training method according to an embodiment of the present disclosure. In this embodiment, a training method Sis applicable to a training model TS. The training method Sincludes steps S~S. In step S, at least one of the communication result CS, the corrected communication result CS', and the feedback (such as positive feedback, negative feedback) for the communication result CS is received from the database. In the embodiment of the present disclosure, the databasemay transmit at least one of the communication result CS, the corrected communication result CS', and the feedback for the communication result CS to the training model TS by an electrical connection or a wireless communication method.

320 In step S, the training model TS determines whether the user is satisfied with the communication result CS. Furthermore, the training model TS determines whether the user is satisfied with the communication result CS according to the feedback for the communication result CS.

121 330 When the training model TS determines that the feedback for the communication result CS is positive, the training model TS may submit the communication result CS to the AI communication toolin step S.

320 340 In step S, when the training model TS determines that the feedback for the communication result CS is negative, the training model TS may determine in step Swhether the classification of the corrected communication result CS' is the first classification or the second classification, thereby further determining whether the user is satisfied with the corrected communication result CS'.

340 121 350 In step S, when the training model TS determines that the corrected communication result CS' is the first classification, it means that the user is satisfied with the corrected communication result CS'. Therefore, the training model TS may submit the corrected communication result CS' classified as the first classification to the AI communication toolin step S.

340 360 2 121 121 2 130 2 In step S, when the training model TS determines that the corrected communication result CS' is classified as the second classification, it means that the user is not satisfied with the corrected communication result CS'. Therefore, the training model TS may correct or edit the corrected communication result CS' in the second classification in step S, thereby generating an edited corrected communication result CS'', and submit the edited corrected communication result CS2'' to the AI communication tool. The AI communication toolsubmits the edited corrected communication result CS'' to the database. In this embodiment, both the corrected communication result CS' and the edited corrected communication result CS'' may serve as training samples for the training model TS.

130 320 340 340 130 In some embodiments, the training model TS receives the communication result CS and the corrected communication result CS' together from the databaseand executes steps Sand Ssimultaneously. In some embodiments, the training model TS may execute only step Saccording to the corrected communication result CS' received from the database.

1 FIG. 5 FIG. 5 FIG. 110 1 3 1 4 110 2 2 4 2 130 3 Please refer toand.illustrates a schematic view of an operation interface according to an embodiment of the present disclosure. In this embodiment, the operation interfaceincludes regions A~Aand buttons B~B. The operation interfacemay display the communication result CS in the region A. If the user is satisfied with the communication result CS displayed in the region A, the user may click the button Bto submit the communication result CS currently displayed on the region Ato the database. Then, the user may click the button Bto terminate the editing of the communication result CS.

2 1 110 4 1 122 122 1 1 1 110 3 If the user is not satisfied with the communication result CS displayed in the region A, the user may input the correction prompt CP in the region Aof the operation interface. Next, the user may click the button Bto submit the correction prompt CP displayed in the region Ato the prompt analysis model. The prompt analysis modelgenerates the answers ANS~ANSn according to the input correction prompt CP and selects the optimized answer ANS according to the similarities SML~SMLn between the correction prompt CP and the answers ANS~ANSn, and corrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS'. The operation interfacedisplays the corrected communication result CS' in the region A.

3 1 3 4 3 130 In this embodiment, when the user is satisfied with the corrected communication result CS' displayed in the region A, the user may click the button Bto mark the corrected communication result CS' currently displayed in the region Aas a satisfactory answer (i.e., provide positive feedback). Then, the user may click the button Bagain to submit the corrected communication result CS' displayed in the region Ato the database.

3 2 3 3 4 1 130 In this embodiment, if the user is not satisfied with the corrected communication result CS' displayed in the region A, the user may click the button Bto mark the corrected communication result CS' currently displayed in the region Aas an unsatisfactory answer (i.e., provide negative feedback). In addition, the user may also edit the corrected communication result CS' displayed in the region A, and click the button Bagain to submit the edited corrected communication result CS'' to the database.

3 Alternatively, when the user has no intention to manually edit the corrected communication result CS', the user may click the button Bto terminate the editing of the corrected communication result CS'.

1 FIG. 6 FIG. 6 FIG. 6 FIG. 1 FIG. 1 FIG. 122 121 110 122 Please refer toand.illustrates a block diagram according to an embodiment of the present disclosure. The operation ofis applicable to a system. The system includes a chatbot CB, a feedback interface FI, a prompt analysis model, and a training model TS. The chatbot CB may include the AI communication toolas shown infor implementation. The feedback interface FI may include the operation interfaceas shown in. The prompt analysis modelincludes an improvement module RSV and an inference module ISV. The chatbot CB may output the corresponding communication result CS according to the data and question provided by the user. When the user is not satisfied with the communication result CS generated by the chatbot CB, the feedback interface FI submits the communication result CS and the correction prompt CP to the improvement module RSV.

1 1 The improvement module RSV generates the answers ANS~ANSn according to the correction prompt CP and submits the answers ANS~ANSn to the inference module ISV.

1 1 1 1 The inference module ISV compares the content of the answers ANS~ANSn with that of the correction prompt CP to generate the similarities SML~SMLn corresponding to the answers ANS~ANSn. The inference module ISV provides the similarities SML~SMLn to the improvement module RSV.

1 1 The improvement module RSV selects the optimized answer ANS from the answers ANS~ANSn according to the similarities SML~SMLn and corrects the communication result CS according to the optimized answer ANS to generate the corrected communication result CS'. The improvement module RSV submits the corrected communication result CS' to the feedback interface FI. Therefore, the user is able to read the corrected communication result CS' through the feedback interface FI.

1 1 1 In this embodiment, the improvement module RSV may submit the corrected communication result CS' and the answers ANS~ANSn to the training model TS. The training model TS may use the corrected communication result CS' and the answers ANS~ANSn as training samples. In some embodiments, the improvement module RSV may submit the corrected communication result CS' and the answers ANS~ANSn to the training model TS through the feedback interface FI.

122 In this embodiment, the chatbot CB and the feedback interface FI may be disposed in a cloud device or an electronic device, respectively. The prompt analysis modeland the training model TS may be disposed in the same or different equipment (e.g., servers).

In this embodiment, the improvement module RSV and the inference module ISV are implemented by, for example, computing circuits or processors of any form.

In summary, the prompt analysis model of the present disclosure automatically generates at least one answer according to the correction prompt and uses the at least one answer as response samples to train the AI communication tool. In this way, the user does not need to spend so much time to provide multiple response samples. The training method is able to obtain a large amount of response samples. In addition, the prompt analysis model selects an optimized answer from the at least one answer and corrects the communication result according to the optimized answer to generate a corrected communication result. The training method may also generate a corrected communication result that is the closest to the correction prompt.

Although the present disclosure has been disclosed above with embodiments, they are not intended to limit the present disclosure. Any person having ordinary knowledge in the technical field may make minor modifications and refinements without departing from the spirit and scope of the present disclosure. Therefore, the scope to be protected by the present disclosure shall be defined by the appended claims.